Your dashboard says stores win. Does the data agree?
A retail satisfaction dashboard shows 80.7% for stores and 69.8% for online, so the obvious read is that stores are 10.9 points better. Break the same 6,050 surveyed customers apart by product category and online is ahead in every single one. This template lays out the counts, shows the reversal, and runs a regression that puts a number on the real channel effect.
Analytics Advanced Regression free
After you install, this is the model to open.
Simpson's Paradox Detector
- In your spreadsheet, click the Sortia icon in the strip of icons down the right-hand edge. No strip? Click the arrow at the bottom-right to open it. You can also use Extensions, then Sortia, then Open Sortia.
- Click Start from a template and put that name in the search box.
- Pick the card with that name and click Load this template. It arrives on a new tab with real numbers already in it.
The answer
One click runs the regression. The sheet already rebuilds the dashboard number, the six per-category comparisons and the mix-adjusted gap, so you can see the reversal before the statistics arrive.
- Dashboard gap
- 10.9 points, in-store ahead
- Categories where online wins
- 6 of 6 product categories
- Mix-adjusted online lead
- +5.7 points
- Complexity effect
- -4.5 points per complexity point
The dashboard is not wrong about the arithmetic, it is wrong about the question. Stores lead by 10.9 points only because two thirds of their surveyed customers bought simple accessories and staples, while 61% of online volume sits in furniture and electronics where nobody is happy. Category by category, online wins all six, by 4.2 to 7.8 points. Reweighted to a common product mix, online is ahead by 5.7 points, and the regression agrees: an online coefficient of +0.057 with a t-statistic near 10.9, alongside -4.5 points of satisfaction for every point of product complexity. Even a plain average of the 24 cell rates favors stores by 4.1 points until complexity enters the model, which is exactly how a mix effect hides.
The model
Twenty-four cells of survey counts, one per product category and channel and region, plus a complexity score that captures how hard each category is to buy and set up.
| Rows in the grid | 24 cells covering 6,050 surveyed customers |
| Outcome (y) | Satisfaction rate per cell, satisfied divided by surveyed |
| Driver 1 | Online channel, 1 for online and 0 for in-store |
| Driver 2 | Product complexity score, 1 simple to 9 install-heavy |
| In-store volume mix | 66% of surveyed sit in the two simplest categories |
| Online volume mix | 61% of surveyed sit in the two most complex categories |
| Category satisfaction spread | 87.8% for accessories in store down to 52.0% for electronics |
Once it is in your sheet
- The model arrives with real numbers in it and runs as it stands, so you can press the button first and understand it second.
- Change the numbers to yours. The sheet marks which cells are inputs and which hold formulas, and most labels carry a note explaining the row.
- Press the run button at the bottom of the panel. It is labeled for the tool you are in, and the result lands on its own tab, with a written reading of it beside the figures.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Next question
- How long will the next batch take?Learning Curve: How Long Will the Next Batch Take?
- What is your average hiding?One Table That Describes Your Numbers
- You guessed min, likely, max. What does history say?Which Distribution Fits Your Data?
- What season length should your forecast use?Does Your Revenue Have a Season?
- Is that vibration random noise or a real rhythm?Find the Hidden Vibration Frequencies
- Can you get the same random numbers back?Reproducible Random Draws, With a Seed
Every model like this one, and the method behind them: Statistics in Google Sheets.